September 2013 - kernelnewbies@kernelnewbies.org

Checking if a Buffer head is valid or not
by Pranay Srivastava Sept. 29, 2013

Sept. 29, 2013

Hi, I've a question regarding how to check if a buffer head is valid. Suppose that I've read in a block using sb_bread and that buffer head hasn't been touched in a while so that it's uptodate. Now suppose that somebody drops the caches from user space. I looked at the code and found the following, Beginning from drop_caches_sysctl_handler --->iterate_supers ---->drop_pagecache_sb --->invalidate_mapping_pages ---->invalidate_inode_page --->invalidate_complete_page --->try_to_release_page In the try_to_release_page there's call to try_to_free_buffers Now since i would be having a pointer to the buffer head which is already uptodate and nothing going on with it this would actually free the buffer. There's a call to kmem_cache_free, so i would be left with a pointer which is invalid [Correct?] Now how do I check if i need to re-read this buffer from disk since the pointer is actually not valid anymore so checking for bh_uptodate doesn't make sense. Reading the buffer from disk always isn't something i'm looking for so any advice on how to check it? Regards, Pranay Kr. Srivastava -- Pranay Srivastava

1 0

Question about tunnels, IPsec and redirect
by Kevin Wilson Sept. 29, 2013

Sept. 29, 2013

Hi, I am looking at this patch: http://lists.openwall.net/netdev/2007/08/24/29 and I cannot understand it. Can somebody please try to explain ? more specifically: Can somebody please give an example of some setup of IPsec tunnel where the ip_rt_send_redirect() method should not be called when the skb->sp is not NULL ? (in other words, why if the SKB is and IPsec SKB, we should not send a redirect in such a case while forwarding a packet; note I am talking about IPv4) Note that the check for skb->sp was changed in recent kernels to skb_sec_path(skb), but it is essentially the same. Regards, Kevin

2 4

Re: can a bottom half be scheduled by another BH ??
by Vishwas Srivastava Sept. 28, 2013

Sept. 28, 2013

Hi Group,**** I am a bit unclear if, I could schedule another BH while I am already running in a BH ? For example, if I am running in a SOFT IRQ, can I schedule a tasklet from there ?? ** ** Thanks,**** Vishwas S**** ** ** On Fri, Sep 27, 2013 at 4:44 PM, Vishwas Srivastava (vishsriv) < vishsriv(a)cisco.com> wrote: > Hi Group,**** > > I am a bit unclear if, I could schedule another BH while I am already > running in a BH ?**** > > For example, if I am running in a SOFT IRQ, can I schedule a tasklet from > there ??**** > > ** ** > > Thanks,**** > > Vishwas S**** > > ** ** > > ** ** >

1 0

help needed to unsubscribe
by Mukesh Yadav Sept. 27, 2013

Sept. 27, 2013

Hi All, I wish to unsubscribe from mailing list. On the following page I clicked on unsubscribe button by entering my email id. It says conformation has been sent. But did not receive any mail regarding conformation. Please help me on this. <http://lists.kernelnewbies.org/mailman/options/kernelnewbies> -- Regards, Mukesh

2 1

Tracing Association code
by Prabhunath G Sept. 25, 2013

Sept. 25, 2013

Hi All, I am tracing the association code in linux-wireless. I understand that association request has to be triggered from station side (userspace-wpa_supplicant) and has to be sent to the AP (userspace-hostapd). In the process of tracing, I found the function ieee80211_assoc(...) in net/mac80211/cfg.c and also I found that a timer is scheduled to execute the timer function. My Questions are 1. Who is initiating the timer function, my guess is it should be the driver & How ? 2. On the AP side how is the association request taken from driver -> mac80211 -> hostapd. Appreciate your help. Thanks, Prabhu

1 0

BSP and the kernel
by arun kumar Sept. 23, 2013

Sept. 23, 2013

Does the BSP has any role to play in the kernel build process. is the Image "Vmlinux" dependent on BSP in any way?

5 4

working of migrate_page (exported) function in mm/migrate.c linux Kernel
by ajay saini Sept. 23, 2013

Sept. 23, 2013

Hi, Anybody knows the working of this function "migrate_page" (not "migrate_pages") defined in mm/migrate.c file (linux 2.6) This function is exported so can be used in a Kernel module. Is it used somewhere ? Can it be used to migrate a page to a different NUMA node from kernel module and How to use it? Is there a function/way in kernel which can be used to migrate page to different NUMA node from Kernel module (not userspace)? Ajay

2 2

Re: working of migrate_page (exported) function in mm/migrate.c linux Kernel
by PMcphoy Sept. 23, 2013

Sept. 23, 2013

Sent from Samsung Mobile -------- Original message -------- From: Valdis.Kletnieks(a)vt.edu Date: 23/09/2013 02:53 (GMT+00:00) To: ajay saini <ajay_saini1006(a)yahoo.co.in> Cc: kernelnewbies(a)kernelnewbies.org Subject: Re: working of migrate_page (exported) function in mm/migrate.c linux Kernel _______________________________________________ Kernelnewbies mailing list Kernelnewbies(a)kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

1 0

how to implement mmap for sysfs device driver
by Niroj Pokhrel Sept. 20, 2013

Sept. 20, 2013

Hi all, I want to implement a device driver with mmap operations. However, I am not using ioctl call. My approach: i. I have created a sysfs bin files (because it support mmap) ii. Now, I have implemented my mmap in my driver and assign this attribute to the sysfs file I have created. However, I am facing the problem. The mmap is being called properly, however mmap_fault my implementation is not being called as a result of which the memory is not mapped. So, when I try to access this file I get segmentation fault. Can you please suggest what I should modify or is there anyother way to do it other than this ?? Thanking you in advance. Thanks and Regards, Niroj Pokhrel

3 2

How to create sysfs file and associate it with file operations like mmap
by Niroj Pokhrel Sept. 18, 2013

Sept. 18, 2013

Hi all, I want to create a sysfs file and associate it with several file opeations like open,close and mmap. But when coming across sysfs file creation I saw that I can create a file using device_create_file(struct device*,struct device_attributes*); The device attributes were initialized using following DEVICE_ATTR(_name,_permission,_show,_store); But I want to associate my file with following operations struct file_operations my_fops { .open = my_open, .close = my_close, .mmap = my_mmap, }; I was able to do this using debugfs_create_file(...,&my_fops); But is there anyway if I want to put this file in sec_class. Thanking you in advance. Thanks and Regards, Niroj Pokhrel

2 1